Electrospun NafionÂ®/Polyphenylsulfone ... - Semantic Scholar

materials Article

Electrospun Nafion®/Polyphenylsulfone Composite Membranes for Regenerative Hydrogen Bromine Fuel Cells Jun Woo Park 1 , Ryszard Wycisk 1 , Peter N. Pintauro 1, *, Venkata Yarlagadda 2 and Trung Van Nguyen 2 1 2

*

Department of Chemical and Biomolecular Engineering, Vanderbilt University, Nashville, TN 37235, USA; [email protected] (J.W.P.); [email protected] (R.W.) Department of Chemical and Petroleum Engineering, University of Kansas, Lawrence, KS 66045, USA; [email protected] (V.Y.); [email protected] (T.V.) Correspondence: [email protected]; Tel.: +1-615-343-3878

Academic Editor: Dong Hong Received: 21 January 2016; Accepted: 18 February 2016; Published: 29 February 2016

Abstract: The regenerative H2 /Br2 -HBr fuel cell, utilizing an oxidant solution of Br2 in aqueous HBr, shows a number of benefits for grid-scale electricity storage. The membrane-electrode assembly, a key component of a fuel cell, contains a proton-conducting membrane, typically based on the perfluorosulfonic acid (PFSA) ionomer. Unfortunately, the high cost of PFSA membranes and their relatively high bromine crossover are serious drawbacks. Nanofiber composite membranes can overcome these limitations. In this work, composite membranes were prepared from electrospun dual-fiber mats containing Nafion® PFSA ionomer for facile proton transport and an uncharged polymer, polyphenylsulfone (PPSU), for mechanical reinforcement, and swelling control. After electrospinning, Nafion/PPSU mats were converted into composite membranes by softening the PPSU fibers, through exposure to chloroform vapor, thus filling the voids between ionomer nanofibers. It was demonstrated that the relative membrane selectivity, referenced to Nafion® 115, increased with increasing PPSU content, e.g., a selectivity of 11 at 25 vol% of Nafion fibers. H2 -Br2 fuel cell power output with a 65 µm thick membrane containing 55 vol% Nafion fibers was somewhat better than that of a 150 µm Nafion® 115 reference, but its cost advantage due to a four-fold decrease in PFSA content and a lower bromine species crossover make it an attractive candidate for use in H2 /Br2 -HBr systems. Keywords: proton conducting membrane; electrospinning; Nafion; polyphenylsulfone; redox flow battery; regenerative fuel cell; hydrogen fuel cell; bromine

1. Introduction Renewable energy sources like wind and solar can be utilized for the generation of a significant amount of electrical energy in the United States, but their intermittent nature is hindering wide-spread implementation. The development of a suitable electrochemical energy storage system might be one solution to the above problem. Additionally, a reliable and efficient energy storage system could help in reducing electrical grid destabilization by intermittent green sources. One such system, which is scalable to the megawatt size, is a regenerative hydrogen-bromine (H2 /Br2 ) fuel cell that utilizes Br2 in aqueous HBr as the oxidant. This system has several advantages over a regenerative H2 /O2 fuel cell, including: (i) fast bromine oxidation/reduction kinetics which translates into low activation over-potential voltage losses, higher round trip efficiencies and a very high power density on discharge (>1.5 W/cm2 versus 0.7 W/cm2 for H2 /O2 system) [1–6]; (ii) negligible mass transfer limitations due to the high solubility of bromine in the hydrobromic acid electrolyte; (iii) low bromine vapor Materials 2016, 9, 143; doi:10.3390/ma9030143

www.mdpi.com/journal/materials

Materials 2016, 9, 143

2 of 15

pressure, which means that the bromine storage unit and the bromine electrode compartment can be operated without pressurization [7]; (iv) efficient operation with inexpensive carbon cathode, in contrast to regenerative H2 /O2 fuel cells where precious metals such as Ru and Ir are required for oxygen evolution (charging) [8,9] but then Ru- and Ir-based electrodes show poor activity during discharge [8,9], which is a serious challenge for development of regenerative H2 /O2 fuel cells; and finally (v) absence of carbon corrosion at the cathode during charging [10,11], which is an important advantage as carbon corrosion is a problem in regenerative H2 /O2 fuel cells. The operation of a regenerative H2 /Br2 fuel cell is quite simple. During charging, hydrobromic acid (HBr) is electrolyzed to hydrogen (H2 ) and bromine (Br2 ) using electrical energy. These products are stored in external tanks until electricity is needed. During discharging, the stored products, H2 and Br2 , are reacted in the fuel cell to produce HBr and electricity. The membrane-electrode assembly (MEA), a key component of the fuel cell, is composed of a polymeric proton-conducting membrane, typically selected from the perfluorosulfonic acid (PFSA) family, which physically separates the hydrogen electrode and the bromine electrode. The membrane prevents electrical shorting while providing pathways for inter-electrode proton transport and minimizing unwanted Br2 and Br3 ´ crossover. Nafion PFSA membranes possess good thermal/mechanical/chemical stability and high proton conductivity, and have already been utilized in hydrogen-bromine fuel cells [7,12–15]. Nafion membranes, however, suffer from high bromine species (Br´ , Br2 , and Br3 ´ ) crossover. The crossover causes significant columbic losses in the cell and degradation of the platinum catalyst on the hydrogen electrode [1–3,13]. Thus for successful deployment of efficient H2 /Br2 fuel cells, a proton conducting membrane with minimal bromine species permeability is needed. An effective Nafion alternative should be based on a highly charged cation-exchange polymer with a high proton conductivity, which would minimize fuel cell energy losses. Unfortunately, highly charged polymers swell excessively in water and aqueous solutions and are usually brittle in the dry state. Swelling reduces the membrane’s mechanical strength and decreases the effective concentration of fixed charges thus reducing both ionic conductivity and bromine species (co-ion) exclusion. Membrane swelling can be controlled by crosslinking the polymer, but this usually exacerbates the dry membrane brittleness problem [16]. Swelling reduction can also be accomplished by blending the charged polymer with a hydrophobic/uncharged polymer, but often the two polymers are so dissimilar that the resultant phase separation negates the benefits of blending [16]. In order to improve mechanical properties and lower the swelling of highly charged polymers, Pintauro and coworkers have developed new electrospinning techniques enabling fabrication of nanofiber composite ion-exchange membranes from dissimilar polymers [17–21]. In particular, Ballengee and Pintauro prepared stable and mechanically robust composite proton-exchange membranes (PEMs) for hydrogen/air fuel cells using a dual-fiber electrospinning [17]. In the present study, a range of nanofiber composite membranes were fabricated and investigated for use in a H2 /Br2 regenerative fuel cell. The membranes were composed of Nafion® perfluorosulfonic acid (PFSA) ionomer for facile proton transport and uncharged polyphenylsulfone (PPSU) for mechanical reinforcement and control of membrane swelling. This paper is an extension of a previously published study on electrospun Nafion/PVDF composite fuel cell membranes [22], where PPSU is an effective reinforcement replacement for PVDF due to its excellent mechanical characteristics which enables greater control of membrane swelling. 2. Materials and Methods 2.1. Electrospinning Nafion/PPSU Dual nanofiber mats of Nafion and PPSU were prepared by simultaneously electrospinning 1100 EW Nafion PFSA containing poly (ethylene oxide) (PEO) carrier polymer and uncharged polyphenylsulfone (PPSU), as reported previously [17]. Nafion and poly (ethylene oxide) (PEO)


3 of 15

solutions were separately prepared by dissolving Nafion powder (prepared by evaporating the solvent Materials 2016, 9, 143 3 of 14 from Liquion 1115, Ion Power, Inc., New Castle, DE, USA) and PEO powder (Sigma-Aldrich, St. Louis, MO, USA, 400 kDa MW) into a mixed solvent of 2:1 weight ratio n-propanol: water. These two 400 kDa MW) into a mixed solvent of 2:1 weight ratio n‐propanol: water. These two solutions were solutions were then combined to form a Nafion/PEO electrospinning solution where PEO constituted then combined to form a Nafion/PEO electrospinning solution where PEO constituted 1 wt% of the 1total polymer content. For the PPSU fibers, a 25 wt% polymer solution was prepared by dissolving wt% of the total polymer content. For the PPSU fibers, a 25 wt% polymer solution was prepared by ® R 5500NT, from Solvay Advanced Polymers, LLC, 63 kDa MW) in a dissolving PPSU(Radel powder (Radel ® R PPSU powder 5500NT, from Solvay Advanced Polymers, LLC, 63 kDa MW) in a 4:1 4:1 weight ratio mixture of n-methyl-2-pyrrolidone (NMP): acetone. weight ratio mixture of n‐methyl‐2‐pyrrolidone (NMP): acetone. All electrospinning experiments were carried out using a custom-built setup, shown in Figure 1, All electrospinning experiments were carried out using a custom‐built setup, shown in Figure 1, consisting of two syringes filled with the two polymer solutions and driven by two syringe pumps, consisting of two syringes filled with the two polymer solutions and driven by two syringe pumps, two high voltage power supplies and a drum collector. The two polymer solutions were electrospun two high voltage power supplies and a drum collector. The two polymer solutions were electrospun simultaneously from two separate spinnerets (stainless steel needles) placed at the opposite sides simultaneously from two separate spinnerets (stainless steel needles) placed at the opposite sides of of a rotating and laterally oscillating drum collector. The Nafion/PEO solution was electrospun a rotating and laterally oscillating drum collector. The Nafion/PEO solution was electrospun at the at the following conditions: 4.16 kV applied voltage between the needle spinneret and the drum following conditions: 4.16 kV applied voltage between the needle spinneret and the drum collector collector surface (drum surface was grounded), 6.5 cm spinneret-to-collector distance, and a 0.2 mL/h surface (drum surface was grounded), 6.5 cm spinneret‐to‐collector distance, and a 0.2 mL/h solution solution flow rate. The solution PPSU solution was electrospun an applied voltage kV,an an 8.0 8.0 cm flow rate. The PPSU was electrospun at an atapplied voltage of of 7.5 7.5 kV, cm spinneret-to-collector distance, and and a a solution solution flow flow rate rate that spinneret‐to‐collector distance, that was was varied varied from from 0.04 0.04 to to 0.15 0.15 mL/h, mL/h, depending on the desired mat composition. All electrospinning experiments were conducted inside a depending on the desired mat composition. All electrospinning experiments were conducted inside Plexiglas chamber at room temperature, with the relative humidity fixed at 35% ˘ 2%. a Plexiglas chamber at room temperature, with the relative humidity fixed at 35% ± 2%.

Figure 1. Schematic of the dual fiber electrospinning setup used in the present study. Figure 1. Schematic of the dual fiber electrospinning setup used in the present study.

2.2. Dual Nanofiber Mat Processing 2.2. Dual Nanofiber Mat Processing The Nafion/PPSU mats were processed as described earlier [17]. The dual fiber mat was first The Nafion/PPSU mats were processed as described earlier [17]. The dual fiber mat was first compressed four four times times (10 (10 s each) °C. The was exposed then exposed to chloroform compressed s each) at at 16 16 kNkN andand 25 ˝25 C. The mat mat was then to chloroform vapor vapor in a container sealed container 16 min which the PPSU itand it to fill the Nafion voids in a sealed for 16 minfor which softened thesoftened PPSU and caused to fillcaused the voids between ˝ C 1 between Nafion fibers. was The then membrane at 70 °C for h 10 and at 140 °C for by 10 PFSA min, fibers. The membrane dried atwas 70 ˝then C fordried 1 h and at 140 for min, followed ˝ followed by PFSA annealing at 150 °C for 2 h under vacuum. This type of membrane will henceforth annealing at 150 C for 2 h under vacuum. This type of membrane will henceforth be denoted as N be denoted as N (fibers)/PPSU. (fibers)/PPSU. Membranes with the inverse structure, which is that of Nafion reinforced by uncharged PPSU Membranes with the inverse structure, which is that of Nafion reinforced by uncharged PPSU nanofibers, was nanofibers, was also also prepared, prepared, in in the the same same manner manner as as described described by by Ballengee Ballengee and and Pintauro Pintauro [17]. [17]. The dual dual fiber fiber mat mat was was densified densified (compressed) (compressed) at at 107 107 kN kN and and 127 127 ˝°C for 30 30 s. s. The mat was was then then The C for The mat annealed at 150 °C in vacuum for 2 h. These membranes are denoted as N/PPSU (fibers). annealed at 150 ˝ C in vacuum for 2 h. These membranes are denoted as N/PPSU (fibers). ® 115 reference) Prior to testing, all membranes (nanofiber composites and a commercial Nafion Prior to testing, all membranes (nanofiber composites and a commercial Nafion® 115 reference) were boiled boiled inin 1 1 sulfuric acid then in deionized water for boiling each boiling to were MM sulfuric acid andand then in deionized water (one (one hour hour for each step) tostep) ensure ˝ ensure full protonation of the sulfonic acid sites. The membranes were stored in deionized water at full protonation of the sulfonic acid sites. The membranes were stored in deionized water at 25 C. 25 °C. 2.3. SEM Microscopy Electrospun mats and freeze‐fractured membrane cross sections were imaged with a Hitachi S‐4200 scanning electron microscope (Hitachi, Hitachinaka, Japan). The dry membrane samples were manually fractured after cooling in liquid nitrogen. The resultant micrographs were analyzed using ImageJ (version 1.38e) [23].


4 of 15

2.3. SEM Microscopy Electrospun mats and freeze-fractured membrane cross sections were imaged with a Hitachi S-4200 scanning electron microscope (Hitachi, Hitachinaka, Japan). The dry membrane samples were manually fractured after cooling in liquid nitrogen. The resultant micrographs were analyzed using ImageJ (version 1.38e) [23]. 2.4. Ion-Exchange Capacity Ion-exchange capacity (IEC) was determined by the standard method of acid exchange and base titration. A membrane sample of known dry weight in the acid form was soaked in 20 mL of 1 M NaCl for 3 h with stirring to exchange cations. The NaCl solution was replaced repeatedly until no H+ was detected in the NaCl rinsing solution. The amount of H+ released into the total NaCl solution volume was measured by titration with 0.01 N NaOH. The IEC of a membrane sample was calculated using Equation (1). IEC pmequiv{gq “ VN{mdry (1) where IEC (mequiv/g) is the ion-exchange capacity (on a dry polymer weight basis), V (mL) is the volume of the NaOH titrating solution, N (mol/L) is the normality of the NaOH titrating solution, and mdry (g) is the dry mass of the membrane. The Nafion volume fraction in a composite membrane was determined from the measured IEC, as per Equation (2). Nafion volume fraction “ pIECcomposite {IEC Na f ion q ˆ pρcomposite {ρ Na f ion q

(2)

where IECcomposite and IEC Na f ion are the measured ion-exchange capacity of a nanofiber composite membrane and a neat Nafion® film (IEC Na f ion = 0.909 mequiv/g), respectively, and ρcomposite and ρ Na f ion are the measured dry density of a nanofiber composite membrane and a neat Nafion® film (ρ Na f ion = 1.87 g/cm3 ). 2.5. Conductivity Measurements In-plane ion conductivity was measured at 25 ˝ C with rectangular pieces cut from the membranes and equilibrated with water or a 2 M HBr solution. An AC impedance method and a BekkTech, 4-electrode cell (Model—BT110, Scribner Associates, Inc. Southern Pines, NC, USA) were employed. The samples equilibrated with water were loaded into the cell and tested while fully immersed in water. Alternatively, the membranes were soaked in 2M HBr solution for 3 h and then loaded quickly into the conductivity cell, after removing excess electrolyte from the membrane surface with filter paper. Resistance was measured at a single frequency of 1 kHz and membrane conductivity was calculated using the following equation, σ “ L{pR ˆ w ˆ δq

(3)

where σ (S/cm) is ion conductivity, L (cm) is the distance between the potential sensing electrodes in the conductivity cell, R (Ω) is the measured resistance, w (cm) is the width of the membrane sample, and δ (cm) is its thickness. 2.6. Membrane Swelling Fully protonated composite membranes and Nafion® 115 were kept in water and in 2 M HBr for at least 24 h at 25 ˝ C prior to a measurement to ensure full equilibration. Then membrane samples were removed from the solutions and quickly wiped with a filter paper to remove surface liquid and their mass and volume were measured. Next the membranes were dried overnight at 60 ˝ C and then


5 of 15

for 2 h at 100 ˝ C, and the mass and volume were re-measured. Gravimetric and volumetric swelling were calculated using Equation (4): Membrane swelling (%) “

xwet ´ xdry ˆ 100 xdry

(4)

where x was either the membrane’s mass or volume. 2.7. Diffusivity Measurements A transient electrochemical breakthrough method [12,24–26] was used to determine the diffusion coefficient of Br2 /Br3 ´ in membrane samples (where Br3 ´ is produced by the following reaction: Br2 + Br´ = Br3 ´ ). A schematic diagram of the two-compartment diffusion cell employed in the Materials 2016, 9, 143 5 of 14 experiment is shown in Figure 2.

Figure 2. Schematic determining membrane membrane Figure 2. Schematic diagram diagram of of the the two two compartment compartment apparatus apparatus for for determining diffusion coefficients. diffusion coefficients.

The downstream compartment contained a platinum mesh counter electrode and a saturated The downstream compartment contained a platinum mesh counter electrode and a saturated calomel reference electrode. The working Pt/C electrode was hot pressed onto the backside of a calomel reference electrode. The working Pt/C electrode was hot pressed onto the backside of a membrane to create a single electrode membrane‐electrode‐assembly (a half MEA). Here a standard membrane to create a single electrode membrane-electrode-assembly (a half MEA). Here a standard decal method was employed where a catalyst ink electrode (0.4 mg 2Pt/cm2) was painted onto a decal method was employed where a catalyst ink electrode (0.4 mg Pt/cm ) was painted onto a Teflon® Teflon® PTFE film and then transferred to the membrane by hot‐pressing at 140 °C and 0.7 MPa for 3 PTFE film and then transferred to the membrane by hot-pressing at 140 ˝ C and 0.7 MPa for 3 min. The min. The electrode was composed of Pt/C catalyst (40% Pt/C from Johnson Matthey), 5 wt% Nafion electrode was composed of Pt/C catalyst (40% Pt/C from Johnson Matthey), 5 wt% Nafion (from a (from a Sigma‐Aldrich Nafion dispersion) and 0.2 wt% glycerol. The final Pt/C: Nafion weight ratio Sigma-Aldrich Nafion dispersion) and 0.2 wt% glycerol. The final Pt/C: Nafion weight ratio was 77:23. was 77:23. After installing the membrane in the cell, both compartments were filled with 2 M HBr and the After installing the membrane in the cell, both compartments were filled with 2 M HBr and the platinum working electrode was polarized to +0.3 V vs. the saturated calomel reference electrode. platinum working electrode was polarized to + 0.3 V vs. the saturated calomel reference electrode. Once the current stabilized at 50 µA, the electrolyte in the feed compartment was rapidly drained and Once the current stabilized at 50 μA, the electrolyte in the feed compartment was rapidly drained replaced with a pre-mixed solution of 2 M HBr with 0.14 M Br2 . and replaced with a pre‐mixed solution of 2 M HBr with 0.14 M Br 2. Bromine species (Br2 /Br3 ´ ) were electrochemically reduced as they permeated through the ‐ were electrochemically reduced as they permeated through the Bromine (Br2/Br3 ) compartment membrane intospecies the downstream and the resultant current transient curve was recorded ˝ membrane into the downstream compartment and the resultant current transient curve was recorded using data acquisition software. All experiments were carried out at 25 C with well-stirred solutions. using data acquisition software. All experiments were carried out at 25 °C with well‐stirred solutions. 2.8. Analysis of Breakthrough Curves 2.8. Analysis of Breakthrough Curves The diffusivity of bromine species in the nanofiber composite and commercial Nafion membranes The diffusivity of bromine species in experimental the nanofiber composite and commercial Nafion was determined by matching current vs. time data to a theoretical transient diffusion membranes was determined by matching current vs. time experimental data to a theoretical transient diffusion model based on Fick’s Second Law. The differential equation and appropriate boundary and initial conditions for the diffusion cell experiment are presented below. ∂ ∂

∂ ∂

(5)

C = C0 for x = 0 at t ≥ 0

(6)

C = 0 for x = L at t ≥ 0

(7)


6 of 15

model based on Fick’s Second Law. The differential equation and appropriate boundary and initial conditions for the diffusion cell experiment are presented below. BC B2 C “D 2 Bt Bx

(5)

C “ C0 for x “ 0 at t ě 0

(6)

C “ 0 for x “ L at t ě 0

(7)

C “ 0 for 0 ď x ď L at t ă 0

(8)

where D (cm2 /s) is diffusivity, C0 (mol/L) is the membrane-phase Br2 /Br3 ´ concentration at the upstream (feed compartment) membrane/solution interface, and L (cm) is the membrane thickness. The downstream concentration of Br2 /Br3 ´ is set equal to zero (i.e., all electro-reducible bromine species that diffuse through the membrane react at the downstream (sensing) Pt/C electrode that is attached to the backside of the membrane). The first term of the Laplace transform solution to Equations (5)–(8) is „  Jt 1 1 2 “ 1{2 1{2 exp ´ (9) J8 4τ π τ where Jt is the current density at time t, J8 is the steady-state current density, and τ = Dt/L2 (where KL is the membrane thickness). Also, the bromine species permeability (P, with units of cm2 /s) can be calculated from the measured steady-state bromine species flux and membrane thickness, P“

J8 L nFC b

(10)

where n is the number of electrons involved in the Br2 /Br3 ´ reduction reaction, F is Faraday’s constant, and Cb is the external (bulk) concentration of Br2 /Br3 ´ (in the present study Cb = 0.14 M). 2.9. Fuel Cell Performance A plain carbon paper (SGL Sigracet 10AA) was used as the Br2 electrode and a bi-layer gas diffusion medium consisting of carbon paper (SGL Sigracet 35BC) coated with Pt/C and Nafion binder was used as the hydrogen electrode. The Pt catalyst loading for the hydrogen electrode was approximately 0.5 mg/cm2 . Two membrane-electrode-assemblies (MEAs) were prepared by hotpressing: one with a membrane composed of 57 vol% N (fibers)/PPSU and the other one with commercial Nafion 115 membrane. A 2 M HBr/2 M Br2 electrolyte mixture was fed to the Br2 electrode and H2 gas at 21 kPa was recirculated through the hydrogen electrode. The H2 and HBr/Br2 pump flow rates were 1380 cm3 /min (97.2 A/cm2 equivalent) and 1.5 cm3 /min (4.3 A/cm2 equivalent during discharge), respectively. In addition, liquid water at a flow rate of 0.05 cm3 /min was injected into the H2 side to humidify the H2 gas and facilitate hydration of the Nafion ionomer binder in the hydrogen electrode. The fuel cell experiments were conducted at 25 ˝ C and at 45 ˝ C. 3. Results and Discussion 3.1. Membrane Morphology Figure 3a shows a surface SEM image of a Nafion/PPSU dual-fiber electrospun mat, with the Nafion/PPSU ratio equal to 57/43 vol/vol and an as-spun porosity of about 80 vol%. The Nafion and PPSU nanofibers are distributed uniformly but are visually indistinguishable and the average fiber diameter is 320 nm. The dual fiber mats were processed into dense, defect free membranes via chloroform vapor exposure followed by conditioning in boiling 1 M H2 SO4 and water. The cross-sectional SEM image of the fully processed membrane is shown in Figure 3b (57 vol% N(fibers)/PPSU). No evidence of defects were found in the micrographs, indicating that PPSU fibers

Figure 3a shows a surface SEM image of a Nafion/PPSU dual‐fiber electrospun mat, with the Nafion/PPSU ratio equal to 57/43 vol/vol and an as‐spun porosity of about 80 vol%. The Nafion and PPSU nanofibers are distributed uniformly but are visually indistinguishable and the average fiber diameter is 320 nm. The dual fiber mats were processed into dense, defect free membranes via chloroform vapor exposure followed by conditioning in boiling 1 M H2SO4 and water. 7 of The Materials 2016, 9, 143 15 cross‐sectional SEM image of the fully processed membrane is shown in Figure 3b (57 vol% N(fibers)/PPSU). No evidence of defects were found in the micrographs, indicating that PPSU fibers were properly softened during the densification and solvent exposure step and formed a continuous were properly softened during the densification and solvent exposure step and formed a continuous phase within the membranes. The retention of Nafion nanofibers in the processed N (fibers)/PPSU phase within the membranes. The retention of Nafion nanofibers in the processed N (fibers)/PPSU membranes was confirmed by selectively extracting PPSU. An example of the remaining Nafion membranes was confirmed by selectively extracting PPSU. An example of the remaining Nafion structure is shown in Figure 3c. As can be seen, a well-interconnected Nafion nanofiber network structure is shown in Figure 3c. As can be seen, a well‐interconnected Nafion nanofiber network remains intact after the PPSU removal. remains intact after the PPSU removal.

Figure 3. SEM images of (a) an electrospun Nafion/PPSU dual nanofiber mat; (b) freeze‐fractured Figure 3. SEM images of (a) an electrospun Nafion/PPSU dual nanofiber mat; (b) freeze-fractured cross section of 57 vol% N(fibers)/PPSU; and (c) surface of the Nafion fiber structure after extraction cross section of 57 vol% N(fibers)/PPSU; and (c) surface of the Nafion fiber structure after extraction of of all PPSU with liquid chloroform. Magnification 5,000X. all PPSU with liquid chloroform. Magnification 5,000X.

3.2. Membrane Swelling

Gravimetric and volumetric swelling of fully protonated composite membranes and Nafion 115 were determined at 25 ˝ C in both water and in 2 M HBr. The results are listed in Table 1. The most straightforward conclusion is that swelling of the composite membranes, whether gravimetric or volumetric, was reduced, as compared to Nafion 115. Additionally, as expected, the swelling decreased with increasing uncharged polymer (PPSU) content. For example, at 50 vol% Nafion, the electrospun composite membrane swelled 11 wt% (18 vol%) in water, which was significantly less than the swelling of Nafion 115 (28 wt% and 52 vol%). Similarly, swelling of the composite membranes in 2 M HBr was smaller than that of Nafion 115, with a nonlinear decrease in swelling with increasing PPSU content. For example, at 25 vol% Nafion, the electrospun composite membrane swelled 4 wt% (7 vol%) in 2 M HBr versus 20 wt% (35 vol%) swelling of Nafion 115. In summary, a significant depression in water and HBr solution uptake was observed as a result of embedding Nafion nanofibers in PPSU. This reduction in membrane swelling was important because bromine species crossover should decrease with decreasing membrane swelling. Table 1. Swelling in water and in 2 M HBr at 25 ˝ C of Nafion/PPSU composite membranes and Nafion® 115.

Membrane Nafion®

115 N (fibers)/PPSU 57 vol% Nafion 50 vol% Nafion 44 vol% Nafion 25 vol% Nafion

Mass Swelling (%)

Volume Swelling (%)

Water

2 M HBr

Water

2 M HBr

28 – 14 11 8 4

20 – 13 12 8 4

52 – 20 18 11 7

35 – 15 13 12 7

3.3. Ion Conductivity In-plane ionic conductivity of nanofiber composite membranes equilibrated in water and in 2 M HBr was measured at 25 ˝ C and is plotted versus Nafion volume fraction in Figure 4. The two points at a Nafion volume fraction of 1.0 represent the conductivity of an electrospun pure Nafion membrane, i.e., and electrospun mat and processed membrane with no PPSU fibers, with conductivities of 0.092 S/cm


8 of 15

and 0.125 S/cm in water and 2 M HBr, respectively. The high conductivity values in HBr are due to the presence of absorbed and mobile H+ and Br´ ions, i.e., the increased concentration of charge carriers in the membrane leads to a higher ionic conductivity [27]. Data obtained for the reference Nafion® 115 membrane are also shown for comparison; the values were slightly lower than those for the electrospun pure Nafion film (0.084 S/cm and 0.107 S/cm, for water and 2 M HBr equilibrated samples, respectively), probably due to a lower membrane swelling. A linear relationship between proton conductivity and Nafion volume fraction for the N (fibers)/PPSU membranes equilibrated in water is evident, which is in good agreement with the findings of Ballengee and Pintauro [17]. Surprisingly, the ionic conductivity of the composite membranes equilibrated in 2 M HBr did not follow a linear mixture rule (dashed line in Figure 4), especially when the Nafion volume fraction was < 0.6, where the data points lie on the conductivity curve for water equilibrated samples. This effect could be associated with the greater selectivity (Br´ rejection) of composite membranes with a high content of PPSU; these membranes exhibited much less swelling and thus were more effective in the Donnan exclusion of co-ions (due to an increased concentration of membrane fixed charges). 8 of 14 Materials 2016, 9, 143

Figure 4. 4. In‐plane as as a function of Figure In-planeionic ionicconductivity conductivityof ofNafion Nafionnanofiber nanofibercomposite compositemembranes membranes a function ˝ Nafion volume fraction at 25 °C (○) N(fibers)/PPSU measured in water; (●) N (fibers)/PPSU of Nafion volume fraction at 25 C (#) N(fibers)/PPSU measured in water; ( ) N (fibers)/PPSU measured in 2 M HBr. The conductivity of Nafion measured in 2 M HBr. The conductivity of Nafion®® 115 in water and in 2 M HBr is 0.084 S/cm and 115 in water and in 2 M HBr is 0.084 S/cm and 0.107 S/cm, respectively. 0.107 S/cm, respectively. ‐´ 3.4. Bromine species (Br 3.4. Bromine Species (Br22/Br /Br33) Permeation ) Permeation

A typical fit of J t/J8 ∞ experimental data to the theoretical breakthrough curve is shown in Figure A typical fit of Jt /J experimental data to the theoretical breakthrough curve is shown in Figure 5a ® ® 5a for a Nafion 115 membrane [28]; the only the first data eight data were points were to maintain the for a Nafion 115 membrane [28]; only first eight points fitted to fitted maintain the accuracy accuracy of the approximation provided by the first term of the Laplace transform solution to the of the approximation provided by the first term of the Laplace transform solution to the Fick’s equation ´ the ‐ (Equation (9)) [24]. Based(9)) on [24]. the fitted Br2 /Br diffusion calculated Fick’s equation (Equation Based curve, on the the fitted curve, Br2/Br3coefficient diffusion was coefficient was 3 ´ 6 2 2/s, as reported equal to 1.45 ˆ 10 cm× /s, reported earlier [28]. Similar transient curve curve experiments were calculated equal to 1.45 10−6as cm earlier [28]. Similar transient experiments performed with selected electrospun membranes. As anAs example, the fitthe of experimental vs. time were performed with selected electrospun membranes. an example, fit of experimental vs. data to the theoretically predicted breakthrough curve is shown in Figure 5b for the 57 vol% N time data to the theoretically predicted breakthrough curve is shown in Figure 5b for the 57 vol% N ´ in the (fibers)/PPSU membrane. The The resultant species (Br(Br 2 /Br 3 3‐) ) in (fibers)/PPSU membrane. resultant diffusion diffusion coefficient coefficient ofof bromine bromine species the 2/Br 8 cm2 /s, significantly lower than the value of 1.45 ˆ 10´6 cm2 /s composite membrane was 7.79 ˆ 10´ −8 cm 2/s, significantly lower than the value of 1.45 × 10‐6 cm2/s for composite membrane was 7.79 × 10 ® 115. for Nafion ® 115. Nafion

calculated equal to 1.45 × 10−6 cm2/s, as reported earlier [28]. Similar transient curve experiments were performed with selected electrospun membranes. As an example, the fit of experimental vs. time data to the theoretically predicted breakthrough curve is shown in Figure 5b for the 57 vol% N (fibers)/PPSU membrane. The resultant diffusion coefficient of bromine species (Br2/Br3‐) in the composite membrane was 7.79 × 10−8 cm2/s, significantly lower than the value of 1.45 × 10‐6 cm2/s for Materials 2016, 9, 143 9 of 15 Nafion® 115.

Figure 5.5. Fit Fit ofof experimental experimental data data to to the the transient transient breakthrough 115; and Figure breakthroughmodel modelfor: for:(a) (a)Nafion Nafion ® 115; and (b) 57 vol% N (fibers)/PPSU nanofiber composite membrane. Figure 5a adapted with permission (b) 57 vol% N (fibers)/PPSU nanofiber composite membrane. Figure 5a adapted with permission from from ECS Transactions, 50 (2) 1217 (2012). Copyright 2012, The Electrochemical Society. ECS Transactions, 50 (2) 1217 (2012). Copyright 2012, The Electrochemical Society. ®

Br2/Br3‐ diffusion coefficients of all nanofiber composite membranes are plotted in Figure 6a as a Br2 /Br3 ´ diffusion coefficients of all nanofiber composite membranes are plotted in Figure 6a function of Nafion volume fraction. An increase in membrane PPSU content led to a significant as a function of Nafion volume fraction. An increase in membrane PPSU content led to a significant 2/s for pure electrospun Nafion to reduction in the Br2/Br3´‐ diffusion coefficient, from 1.90 × 10−6´ cm 6 cm 2 /s for pure electrospun Nafion to reduction −9in the Br /Br diffusion coefficient, from 1.90 ˆ 10 2 3 2.20 × 10´9 cm22/s for the membrane with 75 vol% PPSU. Comparing the above data with the diffusion Materials 2016, 9, 143 9 of 14 2.20 ˆ 10 cm /s for the membrane with 75 vol% PPSU. Comparing the above data with the diffusion coefficient obtained for N (fibers)/PVDF membranes in [18], it can be concluded that replacing the coefficient obtained for N (fibers)/PVDF membranes in [18], it can be concluded that replacing the ‐ PVDF matrix with PPSU led to a 35‐fold reduction in Br diffusion coefficient at 25 vol% Nafion 2/Br3´ PVDF matrix with PPSU led to a 35-fold reduction in Br2 /Br 3 diffusion coefficient at 25 vol% Nafion content. Similarly, there was 15‐fold diffusion coefficient reduction was replaced by content. Similarly, there was 15-fold diffusion coefficient reduction whenwhen PVDFPVDF was replaced by PPSU PPSU at 50 vol% These Nafion. These differences were to attributed to swelling the higher swelling of the N at 50 vol% Nafion. differences were attributed the higher of the N (fibers)/PVDF (fibers)/PVDF membranes in 2 M HBr, as compared to the swelling of N (fibers)/PPSU composites at membranes in 2 M HBr, as compared to the swelling of N (fibers)/PPSU composites at the same the same Nafion content. Nafion content.

(a)

(b)

Figure 6. 6. (a) (a) Diffusion Diffusion coefficient coefficient of of bromine bromine species species (Br (Br and Br´‐) in Nafion nanofiber composite Figure 2 2and Br3 3 ) in Nafion nanofiber composite membranes of different Nafion volume fractions at 25 °C. Diffusion coefficient of Nafion membranes of different Nafion volume fractions at 25 ˝ C. Diffusion coefficient of Nafion®® 115 was 115 was ´6 2 /s; (b) Steady-state permeability of bromine species (Br and Br ´ 1.45 cm2/s; for the the nanofiber 1.45 ˆ × 10 −6 cm (b) Steady‐state permeability of bromine species (Br22 and Br33‐) ) for nanofiber composite membranes as a function of Nafion volume fractions at 25 ˝ C. Steady-state permeability of composite membranes as a function of Nafion volume fractions at 25 °C. Steady‐state permeability of ® 115 was 4.26 ˆ 10−6´6 cm 2 Nafion cm2/s. /s. Nafion® 115 was 4.26 × 10

The experimentally experimentally determined determined steady-state steady‐state Br22/Br membrane permeability is is plotted The /Br3‐3 ´ membrane permeability plottedas asa permeability decreased decreased with with afunction functionof ofNafion Nafionvolume volumefraction fractionin in Figure Figure 6b. 6b. As As expected, expected, the the permeability ´ ‐ increasing same trend asas thethe Br2Br /Br increasing content content of of the the uncharged uncharged polymer polymer following following the the same trend diffusion 2/Br 3 3 diffusion coefficient. The decrease decrease is is most most likely likely related related to to an an increase increase in in tortuosity and decrease coefficient. The tortuosity and decrease of of the the cross-sectional area for diffusion, along with the concurrent swelling reduction, all of which contributed cross‐sectional area for diffusion, along with the concurrent swelling reduction, all of which ‐ to improvement in membrane’s Br2 /Br3 ´ barrier It was noted that both the diffusion contributed to improvement in membrane’s Br2/Brproperties. 3 barrier properties. It was noted that both the diffusion coefficient and permeability were lower for Nafion 115 membrane compared to those values for membrane from electrospun and densified Nafion (without PPSU). This could indicate either a reduced level of crystallinity or the presence of some residual porosity in the processed electrospun films, where the latter could have occurred during extraction of the PEO carrier polymer. The concept of relative selectivity, as proposed by Cussler and coworkers for characterization of


10 of 15

coefficient and permeability were lower for Nafion 115 membrane compared to those values for membrane from electrospun and densified Nafion (without PPSU). This could indicate either a reduced level of crystallinity or the presence of some residual porosity in the processed electrospun films, where the latter could have occurred during extraction of the PEO carrier polymer. The concept of relative selectivity, as proposed by Cussler and coworkers for characterization of direct methanol fuel cell membranes [29], was utilized to characterize the nanofiber composite membranes. Taking Nafion® 115 as the reference, the relative selectivity is defined by Equation (11) [30]. ”κı Relative selectivity “

P

composite membrane

(11)

”κı P

Na f ion 115

The ionic conductivities (κ, S/cm) from Figure 4 and the permeabilities (P, cm2 /s) from Figure 6b were combined to calculate relative membrane selectivity, which is plotted as a function of Nafion volume fraction in Figure 7. It can be seen that as the Nafion content decreased, the membrane selectivity increased nonlinearly, because the drop in bromine species (Br2 /Br3 ´ ) crossover was greater than the decrease in the ionic conductivity. For example, the 25 vol% N (fibers)/PPSU membrane had selectivity of 11.0, which means that bromine species crossover flux would be that much lower compared to Nafion 115, if the thickness of the composite membrane were adjusted so that its area specific resistance matched that of Nafion 115. A 25 vol% N (fibers)/PPSU membrane with a thickness of 2.5 µm had a crossover flux equal to that of Nafion 115 but the area specific resistance (ASR) of Materials 2016, 9, 143 10 of 14 the membrane was about 10-times smaller. Similarly, a 25 vol% N (fibers)/PPSU membrane with a membrane with a thickness of 25 μm would have ASR equal to that of Nafion 115 but the crossover thickness of 25 µm would have ASR equal to that of Nafion 115 but the crossover flux of the membrane flux of the membrane would be an order of magnitude smaller. would be an order of magnitude smaller.

Figure 7. Relative, with respect to Nafion 115, selectivity of the nanofiber composite membranes as a Figure 7. Relative, with respect to Nafion 115, selectivity of the nanofiber composite membranes as a function of Nafion volume fractions at 25 °C. (○) N(fibers)/PPSU; (●) Nafion 115. 115. function of Nafion volume fractions at 25 ˝ C. (#) N(fibers)/PPSU; ( ) Nafion®®

3.5. Nafion Nanofiber Composite Membrane Stability in 2 M HBr‐0.14 M Br 3.5. Nafion Nanofiber Composite Membrane Stability in 2 M HBr-0.14 M Br22 Membrane chemical stability evaluated by measuring ion conductivity, diffusion Membrane chemical stability waswas evaluated by measuring ion conductivity, diffusion coefficient, ˝ coefficient, and steady‐state permeability of membranes immersed in 0.14 M Br 2 in 2 M HBr at 25 °C and steady-state permeability of membranes immersed in 0.14 M Br2 in 2 M HBr at 25 C for 12, 66, for 12, 66, 135, and 183 hours. These tests were performed with a 55 vol% N (fibers)/PPSU 135, and 183 hours. These tests were performed with a 55 vol% N (fibers)/PPSU membrane. It was membrane. It was that the remained ionic conductivity remained constant at this 0.058 S/cm during this found that the ionicfound conductivity constant at 0.058 S/cm during soaking test, which soaking test, which indicated that –SO indicated that –SO3 H ion exchange groups bonded to the carbon-fluorine side chains of Nafion were 3H ion exchange groups bonded to the carbon‐fluorine side stable inof a Nafion brominewere environment [31]. As seen in Figure 8, the[31]. diffusion coefficient and8, thethe steady-state chains stable in a bromine environment As seen in Figure diffusion ´8h 2 /s then permeability forthe the steady‐state membrane decreased afterfor 12 the h and then stabilized at 2.40 ˆ 10 cm and coefficient and permeability membrane decreased after 12 and ´ 8 2 −8 2 −8 2 4.60 ˆ 10 cm /s respectively, which might indicate some kind of short-term structural rearrangement stabilized at 2.40 × 10 cm /s and 4.60 × 10 cm /s respectively, which might indicate some kind of short‐term structural rearrangement (densification) and/or chemical reaction between bromine and PPSU. In addition, no irreversible color change of the membranes was observed.

coefficient, and steady‐state permeability of membranes immersed in 0.14 M Br2 in 2 M HBr at 25 °C for 12, 66, 135, and 183 hours. These tests were performed with a 55 vol% N (fibers)/PPSU membrane. It was found that the ionic conductivity remained constant at 0.058 S/cm during this soaking test, which indicated that –SO3H ion exchange groups bonded to the carbon‐fluorine side chains of Nafion were stable in a bromine environment [31]. As seen in Figure 8, the diffusion Materials 2016, 9, 143 11 of 15 coefficient and the steady‐state permeability for the membrane decreased after 12 h and then stabilized at 2.40 × 10−8 cm2/s and 4.60 × 10−8 cm2/s respectively, which might indicate some kind of short‐term structural rearrangement (densification) and/or chemical reaction between bromine and (densification) and/or chemical reaction between bromine and PPSU. In addition, no irreversible color PPSU. In addition, no irreversible color change of the membranes was observed. change of the membranes was observed.

´‐ in N(fibers)/PPSU Figure 8.8. Diffusion Diffusion coefficient coefficient ( (●); and steady‐state permeability N(fibers)/PPSU Figure ); and steady-state permeability (#)(○) of of Br2Br /Br 2/Br 3 3 in nanofiber composite membrane as a function of soak time in a solution of 0.14M Br nanofiber composite membrane as a function of soak time in a solution of 0.142 in 2M HBr at 25 °C. M Br2 in 2 M HBr at 25 ˝ C.

3.6. Bromine species (Br2/Br3‐) Permeation for the Two Complementary Nanofiber Composite Membrane Structures 3.6. Bromine Species (Br2 /Br3 ´ ) Permeation for the Two Complementary Nanofiber Composite Figure 9 shows freeze‐fractured SEM cross section images of N/PPSU (fibers) (Figure 9a) and N Membrane Structures (fibers)/PPSU (Figure 9b) composite membranes, where both membranes were of the same Figure 9 shows freeze-fractured SEM cross section images of N/PPSU (fibers) (Figure 9a) and N composition (57 vol% Nafion and 43 vol% PPSU). The presence of nanofibers, either PPSU (Figure 9a) (fibers)/PPSU (Figure 9b) composite membranes, where both membranes were of the same composition or Nafion (Figure 9b), embedded within a continuous matrix of the second membrane component Materials 2016, 9, 143 11 of 14 (57 vol% Nafion and 43 vol% PPSU). The presence of nanofibers, either PPSU (Figure 9a) or Nafion (Figure 9b), embedded within a continuous matrix of the second membrane component (Nafion in (Nafion in Figure 9a, or PPSU in Figure 9b) is evident. Ionic conductivity and transport properties Figure 9a, or PPSU in Figure 9b) is evident. Ionic conductivity and transport properties for both the for “normal” both the “normal” structure PPSU (a reinforcing matrix Nafion fibers) and the structure (a reinforcing matrix andPPSU Nafion fibers)and and the complementary/inverse complementary/inverse structure (a membrane with reinforcing PPSU fibers and a Nafion matrix) structure (a membrane with reinforcing PPSU fibers and a Nafion matrix) are summarized in Table 2. are summarized in Table 2. As expected based on earlier studies [17,22], the composite membranes As expected based on earlier studies [17,22], the composite membranes with PPSU nanofibers exhibited with PPSU diffusion nanofibers exhibited higher diffusion coefficient and steady‐state permeability, a higher coefficient anda steady-state permeability, as compared to those measured for the as compared to those measured for the membrane with Nafion nanofibers of the same composition. membrane with Nafion nanofibers of the same composition. A similar finding was obtained for A similar finding was fuel obtained for hydrogen/bromine cell membranes composited of N hydrogen/bromine cell membranes composited of fuel N (fibers)/PVDF and N/PVDF (fibers) [22]. ‐ ´ (fibers)/PVDF and N/PVDF (fibers) [22]. The difference in Br /Br permeability between the two 2 3 The difference in Br2 /Br3 permeability between the two membrane morphologies was associated membrane morphologies was which associated with differences in swelling, not change the with differences in swelling, did not change the proton transport which rate butdid affected the diffusivity proton transport rate but affected the diffusivity of the much bulkier bromine species. of the much bulkier bromine species.

Figure 9. SEM micrographs of the cross‐sections of two 57 57 vol% Nafion/PPSU membranes with Figure 9. SEM micrographs of the cross-sections of two vol% Nafion/PPSU membranes with complementary morphologies: (a) N/PPSU(fibers)–PPSU nanofibers embedded in a Nafion matrix; complementary morphologies: (a) N/PPSU(fibers)–PPSU nanofibers embedded in a Nafion matrix; and (b) N(fibers)/PPSU–Nafion nanofibers embedded in a PPSU matrix. Magnification is 5,000X. and (b) N(fibers)/PPSU–Nafion nanofibers embedded in a PPSU matrix. Magnification is 5,000X. Table 2. Transport properties of the two nanofiber composite membrane structures, with the same Nafion content. The measured properties of a Nafion 115 reference film are also listed. Structure N/PPSU Nafion fibers 57 vol%

Ion Conductivity (mS/cm) – 65

Diffusion Coefficient (cm2/s) – 7.36 × 10−8

Permeability (cm2/s) – 1.22 × 10−7

Relative Selectivity – 2.1


12 of 15

Table 2. Transport properties of the two nanofiber composite membrane structures, with the same Nafion content. The measured properties of a Nafion 115 reference film are also listed. Structure

Ion Conductivity (mS/cm)

Diffusion Coefficient (cm2 /s)

Permeability (cm2 /s)

Relative Selectivity

N/PPSU

–

–

–

–

57 vol% Nafion Nafion 115

Nafion fibers PPSU fibers

65 65

10´8

7.36 ˆ 1.41 ˆ 10´7

10´7

1.22 ˆ 2.11 ˆ 10´7

2.1 1.2

107

1.45 ˆ 10´6

4.26 ˆ 10´7

1.0

3.7. H2 -Br2 Regenerative Fuel Cell Performance The H2 -Br2 regenerative fuel cell experiments (charging and discharging) were performed at 25 ˝ C and 45 ˝ C with an electrospun dual fiber 55 vol% N (fibers)/PPSU nanofiber composite membrane and with Nafion 115, which served as a reference. The resultant current–voltage curves are shown Materials 2016, 9, 143 12 of 14 in Figure 10. The first thing that can be noticed is the absence, in the discharge curves, of a vertical drop in the cell voltage at low current density as observed in a typical H2 /O2 fuel cell. This is the Utilization of thinner membranes would lead to even better power output however the price result of fast kinetics at both the negative (H2 ) and the positive (Br2 ) electrode. The kinetic losses would be an increase in bromine species (Br2, Br‐, and Br3‐) crossover, resulting in poisoning and contribute only a small fraction of the overall voltage loss induring a H2 /Br cell,interruption which is in at contrast corrosion of hydrogen electrode catalyst (Pt), particularly H22 fuel supply cell to startup/shutdown and during charging [33]. the significant losses at the O2 electrode due to the sluggish ORR kinetics. Therefore, in H /Br2 Thus, the fuel cell lifetime would be reduced if a 2too fuel cell, the performance is limited mainly by the ohmic resistance in the system (predominantly thin membrane were used. A more extensive discussion/analysis of H2‐Br2 fuel cell performance membrane) and bromine species crossover [32]. with dual‐fiber N (fibers)/PPSU membranes can be found in [34].

˝ (a) and 45 °C ˝ (b) with a 55 vol% N Figure 2/Br2 regenerative fuel cell performance at 25 °C Figure 10.10. HH 2 /Br2 regenerative fuel cell performance at 25 C (a) and 45 C (b) with a 55 vol% N (fibers)/PPSU nanofiber composite membrane, 65 μm in thickness (●); and 115 membrane, (fibers)/PPSU nanofiber composite membrane, 65 µm in thickness ( ); andNafion Nafion 115 membrane, 140 μm in thickness (○). Plots show both charging and discharging curves [34]. Adapted with 140 µm in thickness (#). Plots show both charging and discharging curves [34]. Adapted with permission from J. Electrochem. Soc., 162, F919 (2015). Copyright 2015, The Electrochemical Society. permission from J. Electrochem. Soc., 162, F919 (2015). Copyright 2015, The Electrochemical Society.

4. Conclusions The electrospun membrane thickness was 65 µm and its area-specific resistance (ASR) was equal Nanofiber composite membranes were fabricated from electrospun Nafion/polyphenylsulfone to that of Nafion 115. As shown in Figure 10, the performance of electrospun membrane was somewhat (PPSU) dual‐fiber mats for use in a regenerative hydrogen bromine (H 2/Br2) fuel cell. The resultant better compared to that of Nafion 115 at both 25 ˝ C and 45 ˝ C. At 25 ˝ C the maximum power densities structures consisted of Nafion nanofibers surrounded by uncharged PPSU and PPSU nanofibers 2 2 were 0.32 W/cm and 0.28 W/cm for the composite and Nafion membrane, respectively. At 45 ˝ C, surrounded by Nafion. All membranes were stable in 0.14 M Br 2 ‐2M HBr aqueous solutions. Based the maximum power densities were 0.45 W/cm2 and 0.41 W/cm2 for the composite and Nafion on experimentally measured ionic conductivities and bromine species permeabilities, relative membrane, respectively. The diffusivity and steady-state permeability of bromine species in the selectivities were calculated (where the selectivity is the ratio of conductivity to permeability, as composite membrane were lower than those of commercial Nafion 115, but there was no significant compared to the same ratio of a Nafion 115 reference). Five important conclusions can be made difference in the open circuit voltage (OCV) for the two membranes. based on the experimental results: (1) composite nanofiber membranes had better selectivities than Utilization of thinner would lead containing to even better power output however price Nafion 115, e.g., 2.5 and membranes 11.0 for the membranes 50 vol% and 25 vol% Nafion the fibers ´ ´ would be an increase in bromine species (Br2 , Br , and Br3 ) crossover, resulting in poisoning and embedded in PPSU matrix, respectively; (2) when the PPSU content of a composite membrane was corrosion of there hydrogen (Pt), particularly during interruption at cell 2 supply increased, was a electrode decrease catalyst in ionic conductivity, but an even H greater reduction in bromine startup/shutdown and during charging [33]. Thus, the fuel cell lifetime would be reduced if a too thin species permeability, so the relative selectivity increased with increasing uncharged PPSU polymer content; (3) composite membranes with Nafion nanofibers embedded in PPSU matrix had a lower bromine species permeability, as compared to membranes of similar Nafion content where PPSU nanofibers were embedded in a Nafion matrix; (4) composite membranes with Nafion nanofibers and PPSU matrix had a lower bromine species permeability as compared to previously reported nanofiber composite membranes where Nafion fibers were embedded in a PVDF matrix (where both


13 of 15

membrane were used. A more extensive discussion/analysis of H2 -Br2 fuel cell performance with dual-fiber N (fibers)/PPSU membranes can be found in [34]. 4. Conclusions Nanofiber composite membranes were fabricated from electrospun Nafion/polyphenylsulfone (PPSU) dual-fiber mats for use in a regenerative hydrogen bromine (H2 /Br2 ) fuel cell. The resultant structures consisted of Nafion nanofibers surrounded by uncharged PPSU and PPSU nanofibers surrounded by Nafion. All membranes were stable in 0.14 M Br2 -2M HBr aqueous solutions. Based on experimentally measured ionic conductivities and bromine species permeabilities, relative selectivities were calculated (where the selectivity is the ratio of conductivity to permeability, as compared to the same ratio of a Nafion 115 reference). Five important conclusions can be made based on the experimental results: (1) composite nanofiber membranes had better selectivities than Nafion 115, e.g., 2.5 and 11.0 for the membranes containing 50 vol% and 25 vol% Nafion fibers embedded in PPSU matrix, respectively; (2) when the PPSU content of a composite membrane was increased, there was a decrease in ionic conductivity, but an even greater reduction in bromine species permeability, so the relative selectivity increased with increasing uncharged PPSU polymer content; (3) composite membranes with Nafion nanofibers embedded in PPSU matrix had a lower bromine species permeability, as compared to membranes of similar Nafion content where PPSU nanofibers were embedded in a Nafion matrix; (4) composite membranes with Nafion nanofibers and PPSU matrix had a lower bromine species permeability as compared to previously reported nanofiber composite membranes where Nafion fibers were embedded in a PVDF matrix (where both the PPSU and PVDF based films had the same ionic conductivity); and (5) H2 /Br2 charge/discharge regenerative fuel cell experiments at 25 ˝ C and 45 ˝ C showed that the power output with 65 µm thick nanofiber membrane containing 55 vol% Nafion fibers was ~10% higher than that with a 150 µm thick Nafion® 115 reference. Taking into account the lower bromine species crossover of nanofiber composite membranes and significant cost advantage of such films, due to their relatively low PFSA content, it can be concluded that electrospun composite Nafion/PPSU membranes are attractive candidates for use in H2 /Br2 grid-scale regenerative fuel cell storage systems. Acknowledgments: The authors thank the National Science Foundation (grant EFRI-1038234) and DOE ARPA-E (grant DE-AR0000262) for financial support of this work. Author Contributions: Peter N. Pintauro, Jun Woo Park, Trung Van Nguyen and Ryszard Wycisk conceived and designed the experiments; Jun Woo Park performed the experiments (except fuel cell testing), which was performed by Venkata Yarlagadda; Jun Woo Park, Venkata Yarlagadda, Ryszard Wycisk, Trung Van Nguyen and Peter N. Pintauro analyzed the data; and Jun Woo Park, Ryszard Wycisk and Peter N. Pintauro wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. The founding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, and in the decision to publish the results.

Abbreviations The following abbreviations are used in this manuscript: ASR MEA N OCV PEO PFSA PPSU PVDF SEM

Area specific resistance Membrane electrode assembly Nafion® Open circuit voltage Polyethylene oxide Perfluorosulfonic acid Polyphenylsulfone Polyvinylidene fluoride Scanning electron microscopy


14 of 15

References 1. 2. 3.

4.

5. 6.

7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Livshits, V.; Ulus, A.; Peled, E. High-power H2 /Br2 fuel cell. Electrochem. Commun. 2006, 8, 1358–1362. [CrossRef] Soloveichik, G.L. Battery Technologies for Large-Scale Stationary Energy Storage. Annu. Rev. Chem. Biomol. 2011, 2, 503–527. [CrossRef] [PubMed] Cho, K.T.; Albertus, P.; Battaglia, V.; Kojic, A.; Srinivasan, V.; Weber, A.Z. Optimization and analysis of high-power hydrogen/bromine-flow batteries for grid-scale energy storage. Energy Technol. 2013, 1, 596–608. [CrossRef] Cho, K.T.; Ridgway, P.; Weber, A.Z.; Haussener, S.; Battaglia, V.; Srinivasan, V. High performance hydrogen/bromine redox flow battery for grid-scale energy storage. J. Electrochem. Soc. 2012, 159, A1806–A1815. [CrossRef] Kreutzer, H.; Yarlagadda, V.; Nguyen, T.V. Performance evaluation of a regenerative hydrogen-bromine fuel cell. J. Electrochem. Soc. 2012, 159, F331–F337. [CrossRef] Lin, G.; Chong, P.Y.; Yarlagadda, V.; Nguyen, T.V.; Wycisk, R.J.; Pintauro, P.N.; Bates, M.; Mukerjee, S.; Tucker, M.C.; Weber, A.Z. Advanced hydrogen-bromine flow batteries with improved efficiency, durability and cost. J. Electrochem. Soc. 2016, 163, A5049–A5056. [CrossRef] Yeo, R.S.; Chin, D.T. A Hydrogen-Bromine Cell for Energy Storage Applications. J. Electrochem. Soc. 1980, 127, 549–555. [CrossRef] Kobayashi, K.; Nagao, M.; Yamamoto, Y.; Heo, P.; Hibinoa, T. Rechargeable PEM Fuel-Cell Batteries Using Porous Carbon Modified with Carbonyl Groups as Anode Materials. J. Electrochem. Soc. 2015, 162, F868–F877. [CrossRef] Nagao, M.; Kobayashi, K.; Yamamoto, Y.; Hibino, T. Rechargeable PEM Fuel-Cell Batteries Using Quinones as Hydrogen Carriers. J. Electrochem. Soc. 2015, 162, F410–F418. [CrossRef] Shena, Y.B.; Harada, T.; Teranishib, S.; Hibinoa, T. A solid-state particulate matter sensor based on electrochemical oxidation of carbon by active oxygen. Sensor Actuat. B-Chem. 2012, 162, 159–165. [CrossRef] Tsuneyama, K.; Tetanishi, S.; Hibino, T.; Nagao, S.; Hirata, H.; Matsumoto, S. Low-temperature hydrocarbon combustion over proton conductor/metal–mixed catalysts. J. Catal. 2010, 273, 59–65. [CrossRef] Baldwin, R.S. Electrochemical Performance and Transport Properties of a Nafion Membrane in a Hydrogen-Bromine Cell Environment; NASA: Washington, DC, USA, 1987; p. 89862. Savinell, R.F.; Fritts, S. Theoretical performance of a hydrogen-bromine rechargeable SPE fuel cell. J. Power Sources 1988, 22, 423–440. [CrossRef] Grot, W.; Rajendran, G. Membranes containing inorganic fillers and membrane and electrode assemblies and electrochemical cells employing same. U.S. Patent 5,919,583, 6 July 1999. Glass, W. Hydrogen-Bromine Fuel Cell; Ionics Inc.: Cambridge, MA, USA, 1964. Kerres, J.A. Blended and cross-linked ionomer membranes for application in membrane fuel cells. Fuel Cells 2005, 5, 230–247. [CrossRef] Ballengee, J.B.; Pintauro, P.N. Composite Fuel Cell Membranes from Dual-Nanofiber Electrospun Mats. Macromolecules 2011, 44, 7307–7314. [CrossRef] Choi, J.; Lee, K.M.; Wycisk, R.; Pintauro, P.N.; Mather, P.T. Nanofiber network ion-exchange membranes. Macromolecules 2008, 41, 4569–4572. [CrossRef] Choi, J.; Lee, K.M.; Wycisk, R.; Pintauro, P.N.; Mather, P.T. Nanofiber composite membranes with low equivalent weight perfluorosulfonic acid polymers. J. Mater. Chem. 2010, 20, 6282–6290. [CrossRef] Choi, J.; Lee, K.M.; Wycisk, R.; Pintauro, P.N.; Mather, P.T.J. Electrochem. Sulfonated polysulfone/POSS nanofiber composite membranes for PEM fuel cells. J. Eletrochem. Soc. 2010, 157, B914–B919. [CrossRef] Park, A.M.; Pintauro, P.N. Alkaline Fuel Cell Membranes from Electrospun Fiber Mats. Electrochem. Solid State 2012, 15, B27–B30. [CrossRef] Park, J.W.; Wycisk, R.; Pintauro, P.N. Nafion/PVDF Nanofiber Composite Membranes for Regenerative Hydrogen/Bromine Fuel Cells. J. Membrane Sci. 2015, 490, 103–112. [CrossRef] Image J. Available online: http://rsbweb.nih.gov/ij/index.html (accessed on 8 January 2016). Yeo, R.S.; Mcbreen, J. Transport properties of Nafion membranes in electrochemically regenerative hydrogen/halogen cells. J. Electrochem. Soc. 1979, 126, 1682–1687. [CrossRef]


25. 26. 27. 28. 29. 30. 31. 32. 33.

34.

15 of 15

Kimble, M.C.; White, R.E.; Tsou, Y.M.; Beaver, R.N. Estimation of the Diffusion Coefficient and Solubility for a Ga s Diffusing Through a Membrane. J. Electrochem. Soc. 1990, 137, 2510–2514. [CrossRef] Yeo, S.C.; Eisenberg, A. Physical properties and supermolecular structure of perfluorinated ion-containing (Nafion) polymers. J. Appl. Polym. Sci. 1977, 21, 875–898. [CrossRef] Kusoglu, A.; Cho, K.T.; Prato, R.A.; Weber, A.Z. Structural and Transport Properties of Nafion in Hydrobromic-Acid Solutions. Solid State Ionics 2013, 252, 68–74. [CrossRef] Park, J.W.; Wycisk, R.; Pintauro, P.N. Membranes for a Regenerative H2 /Br2 Fuel Cell. ECS Transactions 2012, 50, 1217–1231. [CrossRef] Libby, B.; Smyrl, W.H.; Cussler, E.L. Composite membranes for direct methanol fuel cells. Electrochem. Solid State 2001, 4, A197–A199. [CrossRef] Lin, J.; Lee, J.; Kellner, M.; Wycisk, R.; Pintauro, P. Nafion-fluorinated ethylene-propylene resin membrane blends for direct methanol fuel cells. J. Electrochem. Soc. 2006, 153, A1325–A1331. [CrossRef] Morrissey, P.; Vesely, D.; Cooley, G. Stability of sulphonate type membranes in aqueous bromine/bromide environments. J. Membr. Sci. 2005, 247, 169–178. [CrossRef] Tucker, M.C.; Cho, K.T.; Spingler, F.B.; Weber, A.Z.; Lin, G. Impact of membrane characteristics on the performance and cycling of the Br2 -H2 redox flow cell. J. Power Sources 2015, 284, 212–221. [CrossRef] Cho, K.T.; Tucker, M.C.; Ding, M.; Ridgway, P.; Battaglia, V.S.; Srinivasan, V.; Weber, A.Z. Cyclic Performance Analysis of Hydrogen/Bromine Flow Batteries for Grid-Scale Energy Storage. Chem. Plus. Chem. 2015, 80, 402–411. [CrossRef] Yarlagadda, V.; Dowd, R.P.; Park, J.W.; Pintauro, P.N.; Nguyen, T.V. A comprehensive study of an acid-based reversible H2 -Br2 fuel cell system. J. Electrochem. Soc. 2015, 162, F919–F926. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons by Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).